For multicentre studies, procedures and timing should be outlined for central collection, evaluation, and reporting of pooled harms data.
PGB/gabapentin and TCAs/selective serotonin-norepinephrine reuptake inhibitors (SNRIs) were considered essentially equivalent as drugs of first choice based on pooled NNT and number needed to harm (NNH) data from available RCTs. 29 SNRIs show promise as potential therapies for other types of neuropathic pain, but clinical trial data to support their use for PHN are not yet available.
The pooled analysis of mortality showed neither benefit nor harm, but trials with a low risk of bias suggested an excess mortality of 11%.
In addition, they should detect changes in health or treatment use from baseline such that within-trial or pooled statistical analyses could establish signals for, or evidence of harm or poor tolerability, or robust evidence for the absence of harm within the study population.
The pooled analysis of overall mortality did not display more harm with HES, which was not entirely the same as previous trials [ 5- 7] and meta-analyses [ 8, 10, 48, 49].
The estimated effects on mortality, clinically suspected VAP and microbiologically confirmed VAP were very uncertain, because the 95% CIs around the pooled effects included no effect, appreciable benefit and appreciable harm.
We estimated the number needed to treat to harm per year (NNH) (and 95% confidence interval) by applying the pooled odds ratio from the meta-analysis to the annual rate of the event in different populations.
Relative benefit (or risk) was calculated using a fixed effects model [ 19] with no statistically significant difference between treatments assumed when the 95% confidence intervals included 1. Number-needed-to-treat (or harm) was calculated using the method of Cook and Sackett [ 20] using pooled observations.
Number-needed-to-treat (or harm) was calculated by the method of Cook and Sackett [ 22], using the pooled number of observations.
